# NornicDB v1.0.11 Release Announcement

[orneryd]

Release Date: January 27, 2026

I am excited to announce NornicDB v1.0.11, a major release that brings enterprise-grade features, significant performance improvements, and enhanced compatibility. This release includes 369 files changed with 48,599 lines of new code, representing one of our largest feature releases to date.

---

🚀 What's New for Users

Qdrant gRPC Compatibility - Use Your Existing Qdrant SDKs

What it means: NornicDB now speaks the same language as Qdrant, allowing you to use existing Qdrant SDKs (Python, Go, Rust, JavaScript, and more) without any code changes. This means zero migration effort when switching from Qdrant to NornicDB.

What you can do:

• Use your existing Qdrant Python client (qdrant-client) directly with NornicDB
• Leverage multi-language SDK support without rewriting client code
• Perform high-throughput vector ingest operations via the native gRPC protocol
• Execute low-latency vector searches using familiar Qdrant APIs
• Seamlessly migrate existing Qdrant applications to NornicDB

Performance benefits:

• Direct protobuf communication (no JSON overhead)
• Native gRPC streaming for batch operations
• Unified vector index shared between Cypher queries and Qdrant gRPC operations

Learn more: Qdrant gRPC Guide | Architecture Details

---

Canonical Graph Ledger - Audit-Grade Mutation Tracking

What it means: Build a canonical truth store with temporal validity, declarative constraints, and a complete audit trail of all graph mutations. Perfect for financial systems, compliance requirements, and data governance.

What you can do:

• Enforce temporal constraints to prevent overlapping validity periods
• Create append-only mutation logs for complete audit trails
• Generate receipts for every transaction with cryptographic hashes
• Query transaction history using temporal procedures
• Implement versioned facts with validity windows
• Build compliance-ready systems with full mutation tracking

Key features:

• Temporal DDL for time-based constraints
• Transaction log queries via CALL db.txlog.* procedures
• Receipt generation with transaction IDs and WAL sequence numbers
• Configurable WAL retention for ledger-grade data retention

Learn more: Canonical Graph Ledger Guide | Temporal Procedures

---

Vector Search Revolution - Faster, Smarter, More Configurable

What it means: A complete refactoring of the vector search system with new candidate generation strategies, configurable HNSW parameters, and a flexible pipeline architecture. Search is now faster, more accurate, and easier to tune for your specific use case.

What you can do:

• Configure HNSW index parameters (M, ef_construction, ef_search) for optimal performance
• Choose from multiple candidate generation strategies:
  • GPU-accelerated K-means clustering for large-scale searches
  • IVF-HNSW hybrid approach for balanced performance
  • Vector ID collapse for deduplication
• Use the new vector pipeline system for complex search workflows
• Leverage ARM64 Metal acceleration on Apple Silicon for 2-4x faster searches
• Benchmark different search strategies to find the optimal configuration

Performance improvements:

• ARM64 Metal acceleration: 2-4x faster HNSW operations on Apple Silicon
• GPU K-means candidate generation for large-scale vector searches
• Improved recall rates with configurable HNSW parameters
• Reduced memory footprint through optimized candidate generation

Learn more: Vector Search Guide | Embedding Search Architecture | HNSW Configuration

---

Storage Enhancements - Better Performance and Reliability

What it means: New storage serialization options, persistent schema management, and enhanced data validation. Your data is stored more efficiently, validated more thoroughly, and recovered more reliably.

What you can do:

• Use msgpack serialization for smaller storage footprint and faster I/O
• Migrate between serialization formats using built-in migration tools
• Benefit from persistent schema storage across restarts
• Enforce property type validation with comprehensive constraint checking
• Recover from WAL corruption automatically with repair utilities
• Configure WAL retention policies for compliance requirements

Performance improvements:

• Msgpack serialization: 20-30% smaller storage footprint
• Faster property validation with optimized type checking
• Improved WAL compaction reducing storage growth
• Enhanced recovery mechanisms reducing downtime

Learn more: Storage Serialization Guide | Property Data Types | WAL Compaction

---

ARM64 SIMD Optimizations - Native Performance on Apple Silicon

What it means: Native ARM64 NEON SIMD instructions for vector operations, providing significant performance improvements on Apple Silicon and other ARM64 platforms.

What you can do:

• Experience 2-4x faster vector similarity calculations on ARM64
• Benefit from native hardware acceleration without external dependencies
• Use optimized cosine, euclidean, and dot-product operations
• Leverage Metal GPU acceleration on macOS for even better performance

Performance improvements:

• 2-4x faster similarity calculations on ARM64 platforms
• Reduced CPU usage for vector operations
• Native hardware acceleration without software fallbacks

Learn more: NEON SIMD Implementation

---

Auto-Embedding Inference - Intelligent Chunking and Generation

What it means: Automatic embedding generation with intelligent text chunking, making it easier to build semantic search capabilities without manual embedding management.

What you can do:

• Automatically generate embeddings for new nodes with intelligent chunking
• Use query chunking for longer queries to improve search accuracy
• Benefit from automatic embedding invalidation and regeneration
• Leverage semantic search without manual embedding configuration

Performance improvements:

• Reduced manual embedding management overhead
• Improved search accuracy through intelligent chunking
• Automatic optimization of embedding dimensions

Learn more: Vector Embeddings Guide | Embedding Search Architecture

---

Replication Enhancements - Secure, High-Performance Multi-Region

What it means: Enhanced replication with TLS encryption, shared-secret authentication, and improved transport reliability. Build secure, multi-region deployments with confidence.

What you can do:

• Enable TLS encryption for replication traffic
• Use shared-secret authentication for secure cluster communication
• Deploy multi-region setups with improved reliability
• Benefit from enhanced standby startup and failover mechanisms
• Monitor replication health with improved diagnostics

Security improvements:

• TLS encryption for all replication traffic
• Shared-secret authentication preventing unauthorized access
• Enhanced transport security with certificate validation

Learn more: Replication Guide

---

Query Autocomplete - Intelligent Cypher Assistance

What it means: Intelligent query autocomplete powered by the Heimdall AI assistant, making it easier to write correct Cypher queries.

What you can do:

• Get intelligent autocomplete suggestions while typing Cypher queries
• Benefit from context-aware query completion
• Reduce query syntax errors with real-time assistance
• Leverage AI-powered query suggestions in the web UI

Learn more: Autocomplete Plugin | Heimdall AI Assistant

---

📊 Performance Improvements Summary

Write Performance

• 30-50% faster writes through Badger optimizations and fast-path queries
• Reduced memory allocations by 20-30% through service caching and reuse
• Faster drop database operations using Badger prefix drop

Search Performance

• 2-4x faster vector searches on ARM64 with Metal acceleration
• Improved recall rates with configurable HNSW parameters
• GPU-accelerated candidate generation for large-scale searches

Storage Efficiency

• 20-30% smaller storage footprint with msgpack serialization
• Faster property validation with optimized type checking
• Improved WAL compaction reducing storage growth

Memory Usage

• Reduced memory allocations through service caching per database
• Configurable cache limits preventing unbounded memory growth
• Optimized candidate generation reducing memory footprint

Learn more: Performance Guides | HTTP/2 Implementation | Single Request Benchmark

---

🔧 Additional Improvements

Authentication

• Auth caching for reduced authentication overhead in high-throughput scenarios
• Improved session management with TTL and invalidation support

Bolt Protocol

• Enhanced PackStream with hash support for message integrity
• Improved chunked message handling for better reliability
• Fixed integer comparison issues for better Neo4j compatibility

Cypher Query Language

• Enhanced CREATE operations with improved return path support
• Better DELETE handling within complex queries
• Improved long query resilience for complex workloads

WAL and Durability

• WAL corruption detection and repair for improved reliability
• Configurable WAL retention for compliance requirements
• Enhanced compaction reducing storage growth

---

📚 Documentation Updates

This release includes extensive new documentation:

• Qdrant gRPC Guide - Complete guide for using Qdrant SDKs with NornicDB
• Canonical Graph Ledger - Building audit-grade graph systems
• Property Data Types - Complete reference for all supported types
• Embedding Search Architecture - Deep dive into vector search internals
• Storage Serialization - Understanding serialization options
• WAL Compaction - Managing write-ahead log retention
• Performance Guides - HTTP/2, optimization options, benchmarking

Browse all documentation: Documentation Index

---

🎯 Migration Notes

For Existing Users

• No breaking changes - All existing functionality remains compatible
• Optional features - New features are opt-in via configuration
• Backward compatible - Existing data and queries work without modification

New Configuration Options

• NORNICDB_QDRANT_GRPC_ENABLED=true - Enable Qdrant gRPC endpoint
• NORNICDB_EMBEDDING_ENABLED=true - Enable auto-embedding inference
• WAL retention configuration for ledger-grade retention
• HNSW configuration parameters for vector search tuning

See: Configuration Guide

---

🙏 Thank You

This release represents weekd of development work focused on enterprise features, performance, and reliability.